What is stratified sampling?

What is stratified sampling?

• A. Sample randomly without replacement within a single group.
• B. Divide population into strata and sample from each stratum.
• C. Sample only from the largest subgroup.
• D. Cluster sampling is used instead.

Answer: (B)

Stratified sampling means dividing the population into internal subgroups (strata) that are similar within each group on an important characteristic, and then sampling from each stratum. The point of this approach is to make sure every subgroup is represented in the sample, which often leads to more precise estimates than sampling from the whole population in one go. In practice, you split the population into strata such as age groups, income levels, or geographic areas, and then take a random sample from each stratum. The number chosen from each stratum can be proportional to its size or allocated equally, depending on the goal. This helps reduce sampling error when different strata would otherwise contribute differently to the overall estimate. Why this option fits best: it explicitly describes dividing the population into strata and sampling from each, which is the essence of stratified sampling. The other approaches describe different methods: sampling randomly within a single group doesn’t ensure representation across distinct subgroups; sampling only from the largest subgroup ignores diversity among the population; cluster sampling involves selecting whole clusters and then sampling within those clusters, which is a different design from stratified sampling.

Stratified sampling means dividing the population into internal subgroups (strata) that are similar within each group on an important characteristic, and then sampling from each stratum. The point of this approach is to make sure every subgroup is represented in the sample, which often leads to more precise estimates than sampling from the whole population in one go.

In practice, you split the population into strata such as age groups, income levels, or geographic areas, and then take a random sample from each stratum. The number chosen from each stratum can be proportional to its size or allocated equally, depending on the goal. This helps reduce sampling error when different strata would otherwise contribute differently to the overall estimate.

Why this option fits best: it explicitly describes dividing the population into strata and sampling from each, which is the essence of stratified sampling.

The other approaches describe different methods: sampling randomly within a single group doesn’t ensure representation across distinct subgroups; sampling only from the largest subgroup ignores diversity among the population; cluster sampling involves selecting whole clusters and then sampling within those clusters, which is a different design from stratified sampling.